WL-HDD wrongly tried to dhcp on "wan" interface
|Reported by:||anonymous||Owned by:||developers|
Reflash brand new Asus WL-HDD with WhiteRussian RC5. Run "nvram-clean.sh" to make sure nvram is tidy.
However after this you still have settings for "wan_*", even though the WL-HDD doesn't have a WAN port:
root@OpenWrt:/# nvram show | grep wan size: 1156 bytes (31612 left) wan_device=vlan1 wan_pppoe_mru=1492 wan_gateway=0.0.0.0 wan_netmask=0.0.0.0 wan_pppoe_demand=0 wan_pppoe_mtu=1492 wan_proto=dhcp wan_pppoe_idletime=60 wan_ipaddr=0.0.0.0 wan_ifname=eth1
The problem is that eth1 clashes with the LAN interface:
lan_ifnames=eth1 eth2 lan_ipaddr=192.168.1.220 lan_proto=static lan_ifname=br0
So what happens is that you get an instance of udhcpc wrongly running on eth1:
root@OpenWrt:/# ps auxwww |grep dhcp 603 root 376 S udhcpc -i eth1 -r 0.0.0.0 -b -p /var/run/eth1.pid -R 1209 root 292 S grep dhcp
and this in turn seems to mess up external IP comms. For example, if you add a default route and nameserver then do "ipkg update", it slows down then stalls completely after about 30K of data has been transferred. But unsetting the wan_proto/wan_ifname values, commit and reboot, and everything works just fine.
I can't tell *exactly* why having this udhcpc running on eth1 is causing this partial loss of comms, but clearly it shouldn't be there anyway.
Change History (3)
comment:1 Changed 10 years ago by mbm
- Milestone set to 1.0-rc6
- Resolution set to fixed
- Status changed from new to closed
comment:2 Changed 10 years ago by anonymous
- Resolution fixed deleted
- Status changed from closed to reopened